Drift: Enhancing LLM Faithfulness in Rationale Generation via Dual-Reward Probabilistic Inference.

Jiazheng Li, Hanqi Yan, Yulan He

Browse the full ACL paper archive.